sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
fengling has joined #m-labs
fengling has quit [Ping timeout: 240 seconds]
sandeepkr_ has joined #m-labs
kuldeep has quit [Ping timeout: 240 seconds]
sandeepkr_ has quit [Read error: Connection reset by peer]
sandeepkr has quit [Ping timeout: 260 seconds]
sandeepkr_ has joined #m-labs
sandeepkr__ has joined #m-labs
sandeepkr_ has quit [Read error: Connection reset by peer]
kuldeep has joined #m-labs
flaviusb has joined #m-labs
fengling has joined #m-labs
sandeepkr__ has quit [Read error: No route to host]
sandeepkr__ has joined #m-labs
attie has quit [Ping timeout: 240 seconds]
sb0 has joined #m-labs
attie has joined #m-labs
<sb0> bb-m-labs, force build artiq
<bb-m-labs> build forced [ETA 20m20s]
<bb-m-labs> I'll give a shout when the build finishes
<bb-m-labs> build #528 of artiq-kc705-nist_clock is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-kc705-nist_clock/builds/528
<sb0> rjo, whitequark: /var/lib/buildbot/.Xilinx/Xilinx.lic is a 30-day evaluation license that supports the new vivado. i'll change it to the non-expiring license that came with the new kc705 later
<sb0> and the board pings now, nice
<sb0> sometimes xilinx actually fix their bugs
<sb0> bb-m-labs, force build artiq-pipistrello-nist_qc1
<bb-m-labs> build forced [ETA 20m01s]
<bb-m-labs> I'll give a shout when the build finishes
<GitHub96> [artiq] sbourdeauducq pushed 1 new commit to master: https://git.io/vKOLR
<GitHub96> artiq/master 375e821 Sebastien Bourdeauducq: test/ctlmgr: keep trying to ping on OSError...
<bb-m-labs> build #248 of artiq-win64-test is complete: Exception [exception python_unittest] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/248
<bb-m-labs> build #802 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/802
<whitequark> sb0: how did you determine it was a xilinx bug?
<sb0> the change to the gateware that caused the problem was minor and unrelated to SoC components involved in ethernet
<sb0> and xilinx software breaks that way, sometimes. i've had to deal with many similar bugs on milkymist...
<sb0> okay so now there's only #499 and possibly #503 before 1.2
<bb-m-labs> build #228 of artiq-pipistrello-nist_qc1 is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-pipistrello-nist_qc1/builds/228
<whitequark> yes. i'm looking into the pipistrello problem
<bb-m-labs> build #529 of artiq-kc705-nist_clock is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-kc705-nist_clock/builds/529
<bb-m-labs> build #249 of artiq-win64-test is complete: Failure [failed python_unittest] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/249 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<bb-m-labs> build #803 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/803 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<sb0> hmm, maybe the OSError was for something more serious
<cr1901_modern> sb0: Is it reproducible on other machines (just wondering)?
<sb0> what, the OSError?
<cr1901_modern> sb0: Yes. I'm assuming it's the same Windows one from yesterday?
<sb0> yes
<sb0> well, you can try running the test
<sb0> the problem appeared when anaconda updated python from 3.5.1 to 3.5.2
<cr1901_modern> I have an environment for python 3.5.1. Does artiq hard-depend on 3.5.2 now, or was that just conda woes?
<sb0> 2.0 does because I have removed the 3.5.1 monkey patches, but you should be able to reproduce it easily on release-1 that still supports 3.5.1
<cr1901_modern> H/o two seconds... git woes
<cr1901_modern> sb0: Okay what do you want me to test (everything's working fine now on my end. Only errors are missing nonessential modules)?
<sb0> are you running python 3.5.2? as i said, this error popped up after the update
<cr1901_modern> "but you should be able to reproduce it easily on release-1 that still supports 3.5.1". Sorry I misunderstood. Upgrading now.
<sb0> sorry s/reproduce/run
FabM has joined #m-labs
<cr1901_modern> sb0: Okay upgraded. No OSErrors yet. Only thing that seems different is that experiments aren't automatically loaded into the GUI when I open artiq_browser.
<sb0> did you run the failing unittest?
<sb0> the only thing that is affected by the problem is the unittest itself afaict, no user-visible parts
<cr1901_modern> sb0: Same test fails, different error. Which probably means I screwed it up somehow.
<cr1901_modern> So don't get excited yet
<cr1901_modern> Gist incoming
<cr1901_modern> sb0: https://gist.github.com/cr1901/a81d7de67684f95c81b503c6d80195fc This test wouldn't require hardware by any chance?
FabM has quit [Quit: ChatZilla 0.9.92 [Firefox 47.0/20160604131506]]
FabM has joined #m-labs
<sb0> that's what my patch did, revert the last commit
<cr1901_modern> sb0: Same error. Was worth a try just in case
sb0 has quit [Quit: Leaving]
sb0 has joined #m-labs
sb0 has quit [Quit: Leaving]
sb0 has joined #m-labs
<cr1901_modern> rjo: While it's on my mind, https://github.com/cr1901/misoc/commit/42889c5f8401abb0163a6c72d28a7124d97a2ac0#diff-9cf476155dc261df09f1e94e61ea9ca6L66 were the old tests leftovers before you finalized the core in artiq? I got them running/some tests fail right now. Probably I'm not shifting values correctly.
<rjo> cr1901_modern: the stuff after the return that you removed was just an unfinished sketch.
<cr1901_modern> rjo: Well, it's a comprehensive test suite. Any objections if I keep it in/modify it to work on the WB core (before surgically removing the WB bus and replacing with CSR bus)?
<cr1901_modern> In any case I think once the CSR core passes the same test suite as the wishbone version, we should be okay.
<GitHub166> [artiq] whitequark pushed 1 new commit to master: https://git.io/vKOEW
<GitHub166> artiq/master c7a5ec9 whitequark: runtime: update ppp code for lwip 2.0.0....
<bb-m-labs> build #530 of artiq-kc705-nist_clock is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-kc705-nist_clock/builds/530
<bb-m-labs> build #250 of artiq-win64-test is complete: Failure [failed python_unittest] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/250 blamelist: whitequark <whitequark@whitequark.org>
<bb-m-labs> build #804 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/804 blamelist: whitequark <whitequark@whitequark.org>
<whitequark> sb0: ok, what's the next priority on artiq?
Sb0000 has joined #m-labs
<Sb0000> whitequark what happens exactly with the attribute writeback of a list of devices?
<Sb0000> Are new objects created?
<whitequark> no
<whitequark> the references inside the list are updated
<whitequark> and the same value as was there is written back
<whitequark> so it's a nop
<Sb0000> What about the attributes of those device objects?
<whitequark> same thing
<whitequark> well, keep in mind attribute writeback is not *recursive*
<Sb0000> Updated separately from the list?
<whitequark> it traverses every quoted mutable object exactly once, in some arbitrary sequence, and writes back all their attributes or elements
<whitequark> yes
<Sb0000> Ok
<Sb0000> Windows slowness would be next imo
<Sb0000> Thanks for fixing all those issues recently
<whitequark> ok, I'll look at Windows now
<whitequark> it's 1.2 or 2.0?
<Sb0000> 2.0 i think... changing the linker has some breakage potential
<Sb0000> Unless there is some other simpler fix
<whitequark> possibly
<whitequark> I mean we can always uh
<whitequark> prefork the linker
<whitequark> this is disgusting as hell but it'll work
<Sb0000> Fork on windows?
<rjo> cr1901_modern: sure.
<whitequark> spawn and make it wait on the input data
<whitequark> same concept
<Sb0000> Or you mean keep a linker process handy
<Sb0000> Well the linker will die after doing its job no?
<Sb0000> So if you run a rapid succession of kernels it wont help
<whitequark> yeah
<Sb0000> Ok doesnt sound like a worthy approach
Sb0000 has quit [Ping timeout: 250 seconds]
fengling has quit [Quit: WeeChat 1.4]
rohitksingh has joined #m-labs
<GitHub90> [artiq] whitequark pushed 1 new commit to master: https://git.io/vKObw
<GitHub90> artiq/master c4dc4e7 whitequark: compiler.testbench.perf_embedding: update for core changes.
<bb-m-labs> build #531 of artiq-kc705-nist_clock is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-kc705-nist_clock/builds/531
<bb-m-labs> build #251 of artiq-win64-test is complete: Failure [failed python_unittest] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/251 blamelist: whitequark <whitequark@whitequark.org>
<bb-m-labs> build #805 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/805 blamelist: whitequark <whitequark@whitequark.org>
FabM has quit [Quit: ChatZilla 0.9.92 [Firefox 47.0/20160604131506]]
sb0 has quit [Read error: Connection reset by peer]
sb0 has joined #m-labs
kristianpaul has quit [Ping timeout: 258 seconds]
<cr1901_modern> rjo: I'm a bit recognizing this, but why does the artiq spi driver write out all bits of data and THEN reads in the desired number of bits (i.e. in series, not parallel) even in full duplex mode?
<cr1901_modern> I'm a bit late* recognizing this
<cr1901_modern> In full duplex/4-wire mode, I would expect an SPI driver to be reading and writing at the same time.
<cr1901_modern> Does ARTIQ only use three wire mode?
<cr1901_modern> sb0: Ping as well ^. Perhaps you can answer this
<cr1901_modern> rjo: Actually nevermind I think I got it; a write to the SPI driver will also do a read. So if for some reason you wanted to read out the data a slave sends while you're writing to the slave, just set the read_length register to 0, so no *extra* bits are shifted in.
sb0 has quit [Quit: Leaving]
rohitksingh has quit [Ping timeout: 246 seconds]
rohitksingh has joined #m-labs
kristianpaul has joined #m-labs
kristianpaul has quit [Changing host]
kristianpaul has joined #m-labs
MiW has quit [Quit: changing servers]
_MiW has joined #m-labs
_MiW is now known as MiW
sandeepkr_ has joined #m-labs
sandeepkr__ has quit [Ping timeout: 276 seconds]
sandeepkr has joined #m-labs
sandeepkr_ has quit [Ping timeout: 258 seconds]
<rjo> cr1901_modern: correct.